Software Engineer by Role, Data Scientist by Impact
With 7+ years of foundational experience in the civil construction industry, I successfully transitioned into data science and software developmentādriven by a passion for AI, machine learning, and intelligent systems.
Successfully transitioned from 7+ years in civil construction management to a thriving career in data science, leveraging analytical thinking and project management skills in the tech industry.
4+ years of experience in Data Science, Business Analytics, Software Development, and Client Engagement
Expert in Python, SQL, and modern AI frameworks including Open AI Agents SDK, LangChain, LangGraph, RAG, and MCP, with proven experience in end-to-end ML pipeline development and production deployment.
Committed to leveraging data science and AI to solve complex business problems, with a focus on delivering actionable insights and intelligent automation solutions that drive measurable business value.
A Streamlit-based IoT support assistant with RAG over product PDFs, FAISS vector store, and MySQL session storage. Enforces selected language (English/Malay), collects feedback, and provides expert contact details. Includes Docker and CI configs.
Source CodeA production-ready agentic AI with smart single-tool selection (Tavily or Wikipedia) powered by Groq Llama 3. Integrates FastAPI backend and Streamlit UI with CI/CD (Jenkins), Docker, and AWS deployment.
Source CodeProduction-ready AI recommender with a modern chat UI built on Flask. Uses RAG with AstraDB vector search and Groq LLM to generate contextual product recommendations. Includes Docker/Kubernetes deployment and Prometheus/Grafana monitoring.
Source CodeA modular, agentic data cleaning pipeline using the orchestrator-worker pattern, powered by LLMs (ChatGroq) and managed via LangChain and Streamlit. Upload your CSV or use the Titanic sample, and the app will profile, plan, and execute data cleaning steps with LLMs, showing a cleaning plan and before/after results. Features modular agents, dynamic planning, and a simple Streamlit UI.
Source CodeA web application that generates personalized bedtime stories for children based on their preferences. Built with Python, Streamlit, and LangChain, this app provides a fun, interactive way to create magical stories for kids.
Source CodeA modular, conversational RAG (Retrieval-Augmented Generation) system built with Streamlit that allows users to upload PDF documents and have interactive conversations about their content.
Source CodeA coding assistant powered by LangChain and Groq, with a modern Gradio UI. This assistant helps you with coding questions, code generation, and detailed explanations.
Source CodeThis project demonstrates the effect of fine-tuning a small language model (DistilGPT-2) on a tiny movie review dataset using Hugging Face Transformers and Datasets. The app provides a Streamlit UI to compare the model's output before and after fine-tuning.
Source CodeAn AI-powered job search application that helps you find your dream opportunities using SerpAPI and Groq LLM, with a beautiful, modern Streamlit interface.
Source CodeA conversational Python application that uses AI, real-time APIs, and the Model Context Protocol (MCP) to provide current weather information and the latest news headlines for any topic, with a modern Streamlit UI.
Source CodeA Gradio-powered web app that lets you interact with your MySQL database using natural language queries, powered by LangChain and Groq LLM. Visualize your data instantly with tables.
Source CodeTransform YouTube videos into Instagram or Medium content using AI agents! This app leverages CrewAI, Streamlit, and quality APIs to generate high-quality, platform-optimized content from any YouTube video link.
Source CodeThe project predicts student score using a web interface built with Python, Flask, and machine learning, deployed on AWS Elastic Beanstalk with CD via AWS CodePipeline using modular coding approach.
Source CodeThe project uses FastAPI, Transformers, and Hugging Face models to create a text summarization system with pipelines for data handling, model training, and evaluation, deployed on AWS with CI/CD via GitHub Actions, Docker using modular coding approach.
Source CodeThe project features an end-to-end machine learning pipeline with MLflow for experiment tracking, covering data handling, model training, and evaluation, plus a Flask app for wine quality prediction, deployed on AWS with CI/CD via GitHub Actions, Docker using modular coding approach.
Source CodeThe project implements a content-based recommendation system for the MyBay shopping app using Flask, SQLAlchemy, and machine learning, with user authentication and product exploration features, added other experimental recommendation systems.
Source CodeThe project analyzes telecom customer churn using SQL, Power BI, and machine learning, featuring data preprocessing, EDA, predictive analysis and re-visualisation, and an interactive Power BI dashboard to gather actionable insights.
Source CodeThe project conducts a comprehensive sales performance analysis for Twiggy Instamart using Power BI to gain insights into KPIs and identify areas for improvement via actionable insights.
Source CodeThe project offers an end-to-end analysis of a global e-commerce dataset (2022-2023) using Python and SQL, providing actionable insights for optimizing business strategies.
Source CodeThe project analyzes given OTT's content catalog to identify trends, popular genres, and opportunities for future content development through actionable insights.
Source CodeThe project uses the Target Brazil's operational information to generate insights for enhancing Target's strategic decision-making, covering customer orders, payments, shipping, product attributes, and demographics.
Source CodeThe package generates synthetic datasets with continuous, categorical, time-series features and noise, ideal for data scientists and engineers needing data for testing, modeling, or privacy concerns.
Source CodeThe project analyzes monthly sales trends, product category performance, and store performance for Starducks Coffee to identify improvement areas and growth opportunities via gathered actionable insights.
Source CodeThe project involves two web scraping tasks to extract product and review data from an eCommerce platform, performed in separate Jupyter notebooks and saved to CSV files for analysis.
Source CodeThe project aims to solve real-world business problems using ML and applied statistics, providing domain experience with industry datasets and developing data-driven insights for decision-making.
Source CodeThe project aims to build predictive machine learning models using supervised learning techniques on industry datasets from healthcare and banking.
Source CodeThe telecom company aims to predict customer churn using historical data to design effective retention strategies. The prediction to be implemented using ensemble methods
Source CodeThe project focuses on solving industry problems using techniques, emphasizing unsupervised learning, synthetic data, clustering, dimensionality reduction, and practical implementations.
Source CodeThe project is about developing a classifier to predict production outcomes in semiconductor manufacturing by identifying the most relevant signals from noisy data.
Source CodeThe project consists of solving two distinct industry-related challenges using basic neural networks which would be regression and classification.
Source CodeThe project is to develop a multi-label text classifier to predict blogger attributes and an interactive chatbot for customer support automation.
Source CodeThe project is to develop two sequential NLP models: one for sentiment analysis of IMDB reviews and another for sarcasm detection in news headlines using a bidirectional LSTM network.
Source CodeThe project is to develop an image classifier to accurately identify plant species from photographs for botanical research.
Source CodeThe project is to develop face detection and recognition models to automate cast and crew information display for a movie streaming application.
Source CodeThe project is to develop a real-time face detection system using a webcam to capture video, detect faces with a pre-trained model, and display results with bounding boxes.
Source CodeThe project is to employ unsupervised learning models for anomaly detection on a custom dataset, complemented by both visual and numerical analyses.
Source CodePython (Via Jupyter Notebook,Google Colab, PyCharm, VS Code, Cursor) and SQL/MS SQL (Via MySQL Workbench,SSMS) for efficient data handling, analysis, modelling and database querying.
Expert proficiency in Pandas, NumPy, SciPy, scikit-learn, MS-Excel, and SQL for comprehensive data transformation, statistical analysis, and deriving actionable insights from complex datasets. Experience with data profiling, feature engineering, and exploratory data analysis.
Expertise in Python visualization libraries (Matplotlib, Seaborn, Plotly) and BI tool (Power BI) for creating interactive dashboards, compelling data stories, and executive-level presentations that drive strategic decision-making.
Combining data visualization with narrative techniques helps communicate insights clearly and drive decision-making, turning numbers into impactful stories.
Comprehensive expertise in supervised/unsupervised learning, ensemble methods, statistical modeling, and advanced ML frameworks. Specialized in building production-ready models for classification, regression, clustering, and recommendation systems with focus on model interpretability and business impact.
Strong data cleaning and preprocessing skills ensure high-quality data for building effective models.
Experience in neural networks, transformers, and frameworks (TensorFlow, PyTorch, Keras, Hugging Face) for computer vision, NLP, and time-series analysis. Experience with transfer learning, fine-tuning, and deploying deep learning models in production environments.
Skills in Flask, FastAPI, Docker, Kubernetes, AWS, GCP, Hugging Face Spaces, Streamlit Cloud, Render, Jenkins, Circle CI, Prometheus, Grafana, SonarQube, Postman, Insomnia
Experience in computer vision (OpenCV, YOLO, CNN architectures), natural language processing (BERT, GPT, transformers), and time-series forecasting (ARIMA, LSTM, Prophet) for solving complex real-world problems across diverse industry domains.
Proficiency in generative AI (LLM) and agentic AI frameworks (LangChain, LangGraph, OpenAI Agents SDK) for building autonomous, context-aware systems. Specialized in prompt engineering, RAG implementation, multi-agent orchestration, and MCP (Model Context Protocol) for advanced automation and intelligent decision-making.
Proficient in front-end development (HTML, CSS, Bootstrap, JavaScript) and modern UI frameworks for creating responsive, user-friendly interfaces. Experience with Streamlit, Gradio, and web application development for data-driven applications, interactive dashboards, and AI-powered tools.
Expert ability to break down complex problems using structured frameworks and write modular, reusable code following software engineering best practices. Specialized in designing scalable architectures, implementing design patterns, and ensuring code quality through testing and documentation.
High emotional intelligence to navigate teamwork, manage stress, and communicate effectively, vital for collaborating with cross-functional teams, influencing stakeholders and presenting data-driven insights clearly and empathetically.
Advanced expertise in crafting precise, context-aware prompts for large language models ensuring high-quality responses and optimal model performance. Specialized in prompt optimization, few-shot learning, chain-of-thought prompting, and developing robust prompt templates for production applications.
Expert proficiency in project management tools (Jira, Confluence) and collaboration platforms (MS Teams, Miro) for agile development and team coordination. Advanced skills in productivity software (MS-Office), design tools (Canva), and cross-platform development across Ubuntu and Windows environments.
Expert proficiency in Git version control, GitHub/GitLab workflows, and maintaining clean, well-documented code repositories. Specialized in collaborative development, CI/CD pipelines, code review processes, and implementing best practices for scalable, maintainable codebases across multiple projects and teams.
Email ID: post.gourang@gmail.com
Contact No.: +91-8971709672